Glutamate is thought to be the major excitatory neurotransmitter in the brain. There are three major subtypes of glutamate receptors in the CNS. These are commonly referred to as kainate, AMPA and N-methyl-D-aspartate (NMDA) receptors (Watkins and Olverman, Trends in Neurosci. 7:265-272 (1987)). NMDA receptors are found in the membranes of virtually every neuron in the brain. NMDA receptors are ligand-gated cation channels that allow Na.sup.+, K.sup.+ and Ca.sup.++ to permeate when they are activated by glutamate or aspartate (non-selective, endogenous agonists) or by NMDA (a selective, synthetic agonist) (Wong and Kemp, Ann. Rev. Pharmacol. Toxicol. 31:401-425 (1991)).
Glutamate alone cannot activate the NMDA receptor. In order to become activated by glutamate, the NMDA receptor channel must first bind glycine at a specific, high affinity glycine binding site which is separate from the glutamate/NMDA binding site on the receptor protein (Johnson and Ascher, Nature 325:329-331 (1987)). Glycine is therefore an obligatory co-agonist at the NMDA receptor/channel complex (Kemp, J. A., et al., Proc. Natl. Acad. Sci. USA 85:6547-6550 (1988)).
Besides the binding sites for glutamate/NMDA and glycine, the NMDA receptor carries a number of other functionally important binding sites. These include binding sites for Mg.sup.++, Zn.sup.++, polyamines, arachidonic acid and phencyclidine (PCP) (Reynolds and Miller, Adv. in Pharmacol. 21:101-126 (1990); Miller, B., et al., Nature 355:722-725 (1992)). The PCP binding site--now commonly referred to as the PCP receptor--is located inside the pore of the ionophore of the NMDA receptor/channel complex (Wong, E. H. F., et al., Proc. Natl. Acad. Sci. USA 83:7104-7108 (1986); Huettner and Bean, Proc. Natl. Acad. Sci. USA 85:1307-1311 (1988); MacDonald, J. F., et al., Neurophysiol. 58:251-266 (1987)). In order for PCP to gain access to the PCP receptor, the channel must first be opened by glutamate and glycine. In the absence of glutamate and glycine, PCP cannot bind to the PCP receptor although some studies have suggested that a small amount of PCP binding can occur even in the absence of glutamate and glycine (Sircar and Zukin, Brain Res. 556:280-284 (1991)). Once PCP binds to the PCP receptor, it blocks ion flux through the open channel. Therefore, PCP is an open channel blocker and a non-competitive glutamate antagonist at the NMDA receptor/channel complex.
One of the most potent and selective drugs that bind to the PCP receptor is the anticonvulsant drug MK-801. This drug has a K.sub.d of approximately 3 nM at the PCP receptor (Wong, E. H. F., et al., Proc. Natl. Acad. Sci. USA 83:7104-7108 (1986)).
Both PCP and MK-801 as well as other PCP receptor ligands [e.g., dextromethorphan, ketamine and N,N,N'-trisubstituted guanidines] have neuroprotective efficacy both in vitro and in vivo (Gill, R., et al., J. Neurosci. 7:3343-3349 (1987); Keana, J. F. W., et al., Proc. Natl. Acad. Sci. USA 86:5631-5635 (1989); Steinberg, G. K., et al., Neuroscience Lett. 89:193-197 1988); Church, J., et al., In: Sigma and Phencyclidine-Like Compounds as Molecular Probes in Biology, Domino and Kamenka, eds., Ann Arbor: NPP Books, pp. 747-756 (1988)). The well-characterized neuroprotective efficacy of these drugs is largely due to their capacity to block excessive Ca .sup.++ influx into neurons through NMDA receptor channels which become over activated by excessive glutamate release in conditions of brain ischemia (e.g., in stroke, cardiac arrest ischemia etc.) (Collins, R. C., Metabol. Br. Dis. 1:231-240 (1986); Collins, R. C., et al., Annals Int. Med. 110:992-1000 (1989)).
However, the therapeutic potential of these PCP receptor drugs as ischemia rescue agents in stroke has been severely hampered by the fact that these drugs have strong PCP-like behavioral side effects (psychotomimetic behavioral effects) which appear to be due to the interaction of these drugs with the PCP receptor (Tricklebank, M. D., et al., Eur. J. Pharmacol. 167:127-135 (1989); Koek, W., et al., J. Pharmacol. Exp. Ther. 245:969 (1989); Willets and Balster, Neuropharmacology 27:1249 (1988)). These PCP-like behavioral side effects appear to have caused the withdrawal of MK801 from clinical development as an ischemia rescue agent. Furthermore, these PCP receptor ligands appear to have considerable abuse potential as demonstrated by the abuse liability of PCP itself.
The PCP-like behavioral effects of the PCP receptor ligands can be demonstrated in animal models: PCP and related PCP receptor ligands cause a behavioral excitation (hyperlocomotion) in rodents (Tricklebank, M. D., et al., Eur. J. Pharmacol. 167:127-135 (1989))and a characteristic katalepsy in pigeons (Koek, W., et al., J. Pharmacol. Exp. Ther. 245:969 (1989); Willets and Balster, Neuropharmacology 27:1249 (1988)); in drug discrimination paradigms, there is a strong correlation between the PCP receptor affinity of these drugs and their potency to induce a PCP-appropriate response behavior (Zukin, S. R., et al., Brain Res. 294:174 (1984); Brady, K. T., et al., Science 215:178 (1982); Tricklebank, M. D., et al., Eur. J. Pharmacol. 141:497 (1987)).
Drugs acting as competitive antagonists at the glutamate binding site of the NMDA receptor such as CGS 19755 and LY274614 also have neuroprotective efficacy because these drugs-like the PCP receptor ligands--can prevent excessive Ca.sup.++ flux through NMDA receptor/channels in ischemia (Boast, C. A., et al., Brain Res. 442:345-348 (1988); Schoepp, D. D., et al., J. Neural. Trans. 85:131-143 (1991)). However, competitive NMDA receptor antagonists also have PCP-like behavioral side-effects in animal models (behavioral excitation, activity in PCP drug discrimination tests) although not as potently as MK-801 and PCP (Tricklebank, M. D., et al., Eur. J. Pharmacol. 167:127-135 (1989)).
An alternate way of inhibiting NMDA receptor channel activation is by using antagonists at the glycine binding site of the NMDA receptor. Since glycine must bind to the glycine site in order for glutamate to effect channel opening (Johnson and Ascher, Nature 325:329-331 (1987); Kemp, J.A., et al., Proc. Natl. Acad. Sci. USA 85:6547-6550 (1988)), a glycine antagonist can completely prevent ion flux through the NMDA receptor channel--even in the presence of a large amount of glutamate.
Recent in vivo microdialysis studies have demonstrated that in the rat focal ischemia model, there is a large increase in glutamate release in the ischemic brain region with no significant increase in glycine release (Globus, M. Y. T., et al., J. Neurochem. 57:470-478 (1991)). Thus, theoretically, glycine antagonists should be very powerful neuroprotective agents, because they can prevent the opening of NMDA channels by glutamate non-competitively and therefore--unlike competitive NMDA antagonists--do not have to overcome the large concentrations of endogenous glutamate that are released in the ischemic brain region.
Furthermore, because glycine antagonists act at neither the glutamate/NMDA nor the PCP binding sites to prevent NMDA channel opening, these drugs might not cause the PCP-like behavioral side effect seen with both PCP receptor ligands and competitive NMDA receptor antagonists (Tricklebank, M. D., et al., Eur. J. Pharmacol. 167:127-135 (1989); Koek, W., et al., J. Pharmacol. Exp. Ther. 245:969 (1989); Willets and Balster, Neuropharmacology 27:1249 (1988); Tricklebank, M. D., et al., Eur. J. Pharmacol. 167:127-135 (1989); Zukin, S. R., et al., Brain Res. 294:174 (1984); Brady, K. T., et al., Science 215:178 (1982); Tricklebank, M. D., et al., Eur. J. Pharmacol. 141:497 (1987)). That glycine antagonists may indeed be devoid of PCP-like behavioral side effects has been suggested by recent studies in which available glycine antagonists were injected directly into the brains of rodents without resulting in PCP-like behaviors (Tricklebank, M. D., et al., Eur. J. Pharmacol. 167:127-135 (1989)).
There have been two major problems which have prevented the development of glycine antagonists as clinically useful neuroprotective agents:
A. Most available glycine antagonists with relatively high receptor binding affinity in vitro such as 7-Cl-kynurenic acid (Kemp, J. A., et al., Proc. Natl. Acad. Sci. USA 85:6547-6550 (1988)), 5,7-dichlorokynurenic acid (McNamara, D., et al., Neuroscience Lett. 120:17-20 (1990)) and indole-2-carboxylic acid (Gray, N. M., et al., J. Med. Chem. 34:1283-1292 (1991)) cannot penetrate the blood/brain barrier and therefore have no utility as therapeutic agents; PA1 B. The only widely available glycine antagonist that sufficiently penetrates the blood/brain barrier-the drug HA-966 (Fletcher and Lodge, Eur. J. Pharmacol. 151:161-162 (1988))--is a partial agonist with micromolar affinity for the glycine binding site. A neuroprotective efficacy for HA-966 in vivo has not been demonstrated nor has it been demonstrated for the other available glycine antagonists because they lack bioavailability in vivo. PA1 lack the PCP-like behavioral side effects common to the PCP-like NMDA channel blockers such as MK801 or to the competitive NMDA receptor antagonists such as CGS19755; PA1 show potent anti-ischemic efficacy because of the non-competitive nature of their glutamate antagonism at the NMDA receptor; PA1 have utility as novel anticonvulsants with fewer side-effects than the PCP-like NMDA channel blockers or the competitive NMDA antagonists; PA1 help in defining the functional significance of the glycine binding site of the NMDA receptor in vivo.
However, one recent success in identifying orally active glycine receptor antagonists was reported by Kulagowski et al., J. Med. Chem. 37:1402-1405 (1994), who disclose that 3-substituted 4-hydroxyquinoline-2(1H)-ones are selective glycine antagonists possessing potent in vivo activity.
A need continues to exist for potent and selective glycine/NMDA antagonists which can penetrate the blood/brain barrier and which: